Twisted and coiled polymer actuator (TCPA) for soft robotic applications

Tsai, Samuel

Permalink

https://hdl.handle.net/2142/129711 Copy

Description

Title

Twisted and coiled polymer actuator (TCPA) for soft robotic applications

Author(s)

Tsai, Samuel

Issue Date

2025-04-23

Director of Research (if dissertation) or Advisor (if thesis)

• Tawfick, Sameh
• King, William P

Doctoral Committee Chair(s)

• Tawfick, Sameh
• King, William P

Committee Member(s)

• Gazzola, Mattia
• Krishnan, Girish

Department of Study

Mechanical Sci & Engineering

Discipline

Mechanical Engineering

Degree Granting Institution

University of Illinois Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Coiled artificial muscle
• soft actuators
• soft robotics

Abstract

Twisted and coiled polymer actuators (TCPAs), often referred to as coiled artificial muscles, are promising candidates for soft and miniature robotics due to their large contractile stroke, high actuation stress, and significant work capacity. They can be made from a variety of off-the-shelf, low-cost polymeric fibers, such as fishing lines and sewing threads, with a relatively simple manufacturing process. They are also scalable in size: They can be made from fibers ranging from 0.02 to 2 mm in diameter. For these reasons, they are suitable for applications in soft grippers, robotic arms, and miniature robots with various modes of locomotion, such as crawling, climbing, and jumping. TCPAs are created by twisting and coiling fiber to form a helical shape. When heat is applied, the fiber untwists, leading to contraction along the coil axis, much like the action of a natural muscle. However, there are currently several knowledge gaps related to both the fundamental understanding of the process-structure-properties of TCPA and the design rules that enable their use in real-world applications. My thesis seeks to fill this knowledge gap and improve the actuation performance of TCPAs for broader applications in soft robotics. In the first chapter of the thesis, I present a novel architecture for a supercoiled TCPA inspired by the hierarchical structure of skeletal muscles. This design aims to mimic the intricate mechanics of muscle tissue and enable effective electric stimulation. The hierarchical structure of muscles results in a nonlinear J-shaped stress-strain curve, initially soft and progressively stiffer at larger strains. This characteristic allows for efficient storage and release of elastic energy with minimal resistance. To replicate this behavior, I developed super- and hypercoiled artificial muscles by twisting and coiling three or nine fishing line fibers, achieving tunable J-shaped stress-strain curves along with effective antagonistic muscle mechanisms. Our fabrication process incorporates a heating wire within the fiber ply, ensuring uniform temperature distribution. A finite element heat transfer simulation model was built to predict muscle temperature and stroke. To showcase the high work capacity of these artificial muscles, I used them in a work-accumulating device, namely a 14.4 g rope-climbing robot. This robot achieves a payload-to-body weight ratio of 14.6. This is the first demonstration of onboard muscle actuation with a load-carrying capacity five times greater than that of robots using electric motors. In the second chapter, I investigate the repeatability of TCPA over 10,000 actuation cycles and derive an empirical law to predict final muscle length changes based on early-cycle behavior. While previous studies have claimed that TCPA stroke remains stable over long-term cycling, the absolute length change of the muscle has not been rigorously studied. To address this, I constructed an isobaric cycling setup that enables rapid hydrothermal actuation through water immersion, enabling 10,000 cycles within 56 hours by varying the muscle temperature between 15 °C and 75 °C at a rate of 20 seconds per cycle. Surprisingly, while the stroke remains unchanged, the final loaded muscle length exhibits diverse creep behaviors—it can remain stable, elongate (creep), or contract (reverse creep). Through extensive experimentation, I derived an empirical law that captures the linear relationship between final muscle length change (ΔL), stroke (α), and passive strain (ε): ε+α=ΔL. Notably, this relation allows the prediction of final length changes after 10,000 cycles using only the first 100 cycles. These findings provide practical design guidelines for ensuring consistent and repeatable actuation in soft robotics applications requiring long-term reliability. In the third chapter, I introduce an innovative pressure-driven coiled artificial muscle featuring a hollow helical geometry with conductive fibers wound around it. This configuration provides anisotropic expansion properties, enabling untwisting under pressure and facilitating phase-transition-driven actuation by Joule heating. When passing an electric current through the conductive fibers, a low-boiling-point liquid inside vaporizes, generating the pressure needed for muscle actuation. The actuation temperature range spans from 25 °C to 45 °C, generating 50% stroke with less than 20 kPa of actuation pressure. The muscle presents a repeatable stroke of 40% for more than 10,000 cycles, demonstrating minimal creep effects due to the thermoset material employed. Through our experiments, I also identified an optimal fiber winding angle and established a linear relationship between fiber untwisting and muscle stroke. This muscle design is particularly suitable for applications that require a low-temperature range or minimal pressure input. Moreover, I use this muscle to study the mechanical energy efficiency of compliant actuators, which is shown to vary with the applied load. Finally, leveraging the scalability of TCPA, I developed a miniature soft jumping robot inspired by the locust leg linkage mechanism. To enhance stability and achieve a single degree of freedom, the design incorporates a parallelogram linkage structure. The fully elastomeric body stores elastic energy before releasing it as kinetic energy for jumping, enabling high energy storage density, miniaturization, and lightweight performance. Using projection additive manufacturing (AM), the robot is fabricated monolithically, eliminating inefficiencies caused by manual assembly. The smallest tested robot measures 7.5 mm in length, weighs 0.216 g, and jumps 60 times its body length. A reduced-order model accurately predicts jumping distances, and onboard coiled artificial muscles drive a latch-triggering mechanism for self-triggered operation. This integration of bio-inspired design, artificial muscles, and advanced AM enables scalable production of high-performing miniature jumping robots for applications in space-constrained environments such as agriculture and maintenance.

Graduation Semester

2025-05

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/129711

Copyright and License Information

Copyright 2025 Samuel Tsai

Owning Collections